Transcript PPT

Populating Quarterly Constant Price
Supply and Use Tables with Seasonally
Adjusted Data
Simon Compton
Methodology Directorate
Office for National Statistics
United Kingdom
Structure of Presentation
• Describe supply and use tables and their use
in the estimation of GDP
• Introduce some key processes of GDP
compilation: balancing, seasonal adjustment,
chain-linking and aggregation
• Consider some of the interactions between
these processes
• Describe development work in the UK
Gross Domestic Product
• Main measure of economic activity and growth
• Can be measured in three ways:
Output = Expenditure = Income or GDP(O) = GDP(E) = GDP(O)
Supply = Demand:
Domestic Production + Imports = Household Expenditure +
Intermediate Consumption + Capital Formation + Exports +
Change in Stocks
GDP(O) = GDP(E)
Domestic Production – Intermediate Consumption = Household
Expenditure + Capital Formation + Change in Stocks + Exports Imports
GDP(I) = wages and salaries + profits
Compilation and Balancing
• Each component can be measured
separately,
• annually, quarterly or monthly
• Results in three different estimates of
GDP
• Data confrontation exercise, known as
‘GDP balancing’ to reconcile these
estimates
• Results in one ‘best’ estimate of GDP
Supply and Use Tables
• Enables data reconciliation at a more
disaggregated level:
– GDP(O) and GDP(E) by product
– GDP(O) and GDP(I) by industry
ONS considering a breakdown into 369
products and 197 industries
Supply and Use Table
Total Output
Total Intermediate
Consumption
Compensation of
Employees
Gross operating
surplus
Taxes on production
less subsidies on
production
Total Output
GVA
Total Use
Exports
Valuables
Inventories
Product Breakdown
GFCF
=
INTERMEDIATE
FINAL
INVESTM EX US
CONSUMPTION BY + CONSUMP +
+
=
ENT
P
E
INDUSTRY
TION
GG
Total Use
NPISH
SU
= PP
LY
Total Supply
Subs. on products
Taxes on products
T&T margins
Product Breakdown
TAXES
I
+ M + AND
SUBS
P
Imports
PRODUCT
OUTPUT BY
INDUSTRY
=
HH
Total Supply
Example of an Input-Output Table
Product
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Agriculture
Forestry
Fishing
Coal extraction
Oil and gas extraction
Metal ores extraction
Other mining and quarrying
Meat processing
Fish and fruit processing
Oils and fats
Dairy products
Grain milling and starch
Animal feed
Bread, biscuits, etc
Sugar
Confectionery
Other food products
Alcoholic beverages
Soft drinks and mineral waters
Tobacco products
etc
etc
Industries' intermediate consumption
1
2
3
4
5
6
7
8
9
Agricul- Forestry Fishing
Coal
Oil and Metal
Other
Meat Fish and
ture
extracgas
ores
mining processfruit
tion
extrac- extracand
ing
processtion
tion
quarrying
ing
3 652
0
0
0
0
3
12
14
1
10
1
2 797
5
1
6
7
6
7
2
39
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-
28
0
1
4
0
0
1
51
0
0
0
2
0
-
204
0
1
1
0
1
0
0
0
0
-
1 516
14
10
1
10
5
0
6
6
-
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-
0
352
1
1
0
1
0
0
0
1
-
3 460
1
6
2 641
42
109
3
295
81
2
0
2
150
-
898
518
1
2
2
745
102
20
40
1
116
1
74
0
-
….
etc
….
etc
Seasonal Adjustment
Wages and salaries
160000
150000
140000
130000
120000
110000
100000
1999Q1
2000Q1
2001Q1
2002Q1
Original series
2003Q1
2004Q1
2005Q1
Seasonally adjusted series
2006Q1
Supply & Use and Seasonal
Adjustment
Supply and Use tables need to be populated
with seasonally adjusted series:
• for data validation
• facilitates the balancing analysis
• focus on final published outputs
Prices
• Adjustment needed to take account of inflation
in order to estimate ‘real’ economic growth
• Achieved by deflation – dividing the ‘current
price’ series by a price index
• Need some way of weighting together different
series – use prices, to generate constant price
series combining different goods and services
• Relative prices change over time
• Chain-linking – generates chain volume
measures BUT not additive
Seasonally Adjusted Series Required for
Constant Price Supply and Use Tables
• Domestic production and intermediate
consumption tables for 197 industries by 369
products in current prices and previous year
prices
• Expenditure components of GDP by product in
current prices and previous year prices
• Income components by industry in current
prices only
Aggregation and Seasonal Adjustment
• Aggregation required for supply and use at a
very low level of disaggregation
• Problematic, particularly for series from
sample surveys where sampling error may
account for most variation
• Practical problems with seasonal adjustment
for a high volume of series
Chain-linking and Seasonal
Adjustment
• Some research indicates greater potential for
bias if seasonal adjustment precedes chainlinking
• Research based on the UK method of chainlinking suggests the opposite (better to
remove seasonality first)
• Seasonality may be reintroduced after chainlinking
• More research needed in this area
• Careful testing for residual seasonality after
chain-linking is advisable
Chain-linking, Aggregation and
Balancing
• Ideally chain-linking should take place at the
most detailed level possible - the level at
which deflation takes place.
• Balancing at a low level of disaggregation
difficult unless an automated procedure
• A more analytical, judgemental approach is
only feasible at a higher level of aggregation
Order of Processing
•
•
•
•
•
•
At the level of deflation express quarterly constant
prices in previous year prices (PYPs)
Aggregate PYPs to the level of seasonal
adjustment
Chain-link quarterly PYPs to Chain Volume
Measures (CVMs).
Seasonally adjust this series
Unchain this series to create seasonally adjusted
current year and previous year prices
Aggregate to the level of publication and chain-link
Transition of Seasonal Adjustment
•
•
•
Current system involves seasonal
adjustment of many thousands of series
This seasonal adjustment incorporates
much knowledge about the series
Strategy for transition involves:
– Use of a sophisticated and robust automatic
option for seasonal adjustment
– A prioritised programme of seasonal adjustment
review, setting up tailored seasonal adjustment
for individual series, where appropriate.
Some complications
•
•
•
•
•
The need for additional published outputs,
sometimes using different classifications or
dimensions to those of the supply and use tables
The interdependent nature of many parts of the
accounts with data being used in two places, or
estimates derived from one component used in
another
Some technical adjustments bridging the gap
between producer (‘basic’) and retail (‘market’)
prices, which are conceptually difficult for seasonal
adjustment and deflation.
Some particular conceptual complexities and data
issues around series for changes in stocks
Likewise for intermediate consumption
Conclusions
•
•
•
•
•
Populating quarterly constant price supply
and use tables with seasonally adjusted
data not easy and requires comprise
Disaggregated level of seasonal adjustment
a particular concern
Design requires careful consideration of
each component of GDP and order of
processing issues
New design not yet implemented in the UK
Many benefits offered by the new framework
with the analytical opportunities it offers.